GigaModular CDU: LiquidStack’s Solution to AI Workloads
As data centre operators face the unprecedented cooling challenge, the industry is working hard to innovate solutions.
It is well known that AI workloads are transforming thermal management requirements across the industry, creating heat densities that traditional air conditioning systems cannot handle economically or efficiently.
As a result, a change is required in data centre design philosophy.
Now, the industry needs cooling systems that can scale rapidly without requiring complete facility redesigns.
Traditional cooling distribution units typically offer fixed capacity, forcing operators to over-provision infrastructure or face constraints as workloads grow. This means that economic pressure to optimise capital expenditure whilst maintaining operational flexibility has driven interest in modular cooling platforms that can expand incrementally.
This is where LiquidStack comes in. A specialist in liquid cooling systems for data centres, the company has launched what it claims is the industry's first modular and scalable coolant distribution unit: The GigaModular CDU.
This platform delivers up to 10 megawatts of cooling capacity through a modular platform designed for pay-as-you-grow installations.
GigaModular CDU: Targeting hyperscale deployments
The GigaModular system uses direct-to-chip liquid cooling, a technology that circulates coolant directly to processors rather than cooling the surrounding air – which handles higher heat densities than traditional air conditioning systems.
“AI will keep pushing thermal output to new extremes and data centres need cooling systems that can be easily deployed, managed and scaled to match heat rejection demands as they rise,” says Joe Capes, CEO of LiquidStack.
LiquidStack has designed the platform specifically for AI high-performance computing and hyperscale data centres.
The modular architecture allows operators to start with 2.5MW of cooling capacity and expand to 10MW as requirements grow. The system can also scale to support hundreds of megawatts across distributed deployments, addressing the needs of large cloud computing providers and AI companies building extensive training facilities.
Operators can deploy the units quickly and scale capacity on demand through centralised controls and simplified service procedures. The pay-as-you-grow model allows companies to match cooling investments with actual thermal loads rather than over-provisioning capacity.
Modular design addressing service and deployment challenges
LiquidStack has also engineered the GigaModular with separate pump and control modules to reduce complexity and component count. High-efficiency IE5 pumps and dual brazed plate heat exchangers deliver scalable performance across different load conditions.
- Fully scalable cooling capacity: Scalable, modular platform supporting single-phase Direct-to-Chip liquid cooling heat loads from 2.5MW to 10MW
- Pump module: High-efficiency IE5 pump and dual BPHx deliver scalable performance; dual 25um strainers protect critical components to ensure reliable operation at scale
- Control module: Centralised design, separate pump and control modules reduce both the number of components and complexity
- Instrumentation kits: Centralised pressure, temperature and EM flow sensors optimise control
- Simplified service access: Serviceable from the front of the unit, with no rear or end access required, allowing the system to be placed against the wall
- Optional configuration: Skid-mounted system with rail and overhead piping pre-installed or shipped as separate cabinets for on-site assembly
The centralised design incorporates pressure, temperature and electromagnetic flow sensors to optimise control systems. Dual 25-micron strainers protect components from contamination during operation.
Additionally, service access has been simplified to allow maintenance from the front of the unit only, eliminating the need for rear or end access – allowing facilities to position units against walls, maximising floor space utilisation.
Operators can choose between skid-mounted systems with pre-installed rail and overhead piping, or separate cabinets for on-site assembly depending on their facility requirements.
“With up to 10MW of cooling capacity at N, N+1, or N+2, the GigaModular is a platform like no other – we designed it to be the only CDU our customers will ever need,” Joe says.
“It future-proofs design selections for direct-to-chip liquid cooling without traditional limits or boundaries.”
How sustainability considerations are driving adoption
The thermal management requirements of AI workloads are reaching levels not previously encountered in commercial data centres.
As hardware manufacturers continue developing more powerful processors, the need for efficient cooling systems becomes increasingly important for operational sustainability.
Liquid cooling systems typically operate with higher efficiency than traditional air conditioning, potentially reducing overall energy consumption in facilities with high thermal densities. This efficiency becomes more significant as AI workloads expand globally.
Therefore, companies deploying AI infrastructure are evaluating cooling technologies that can handle current requirements while accommodating future growth in processing power and thermal output.
LiquidStack recently demonstrated the GigaModular CDU at Datacloud Global Congress in Cannes, France, showing the unique assets of the platform as LiquidStack's response to market demands for scalable thermal management solutions – as AI adoption accelerates across enterprise and hyperscale environments.
“It's not just a CDU, it's an architecture for the future,” the company states.
Explore the latest edition of AI Magazine and be part of the conversation at our global conference series, Tech & AI LIVE.
Discover all our upcoming events and secure your tickets today.
Also sign up to our free weekly newsletter for the latest insights and stories straight into your inbox.
AI Magazine is a BizClik brand